Evaluation of PRISMA products over snow in the Alps and Antarctica

• Main Authors: Mauro, Biagio Di, Cogliati, Sergio, Bohn, Niklas, Traversa, Giacomo, Garzonio, Roberto, Tagliabue, Giulia, Bramati, Gabriele, Cremonese, Edoardo, Julitta, Tommaso, Guanter, Luis, Kokhanovsky, Alexander, Giardino, Claudia, Panigada, Cinzia, Rossini, Micol, Colombo, Roberto
• Format: Other/Unknown Material
• Language: unknown
• Published: Authorea, Inc. 2023
• Subjects: Boa; East Antarctica; Prisma; Antarc*; Antarctica; Ice Shelf
• Online Access: http://dx.doi.org/10.22541/essoar.170365189.95440911/v1

PRISMA is a hyperspectral satellite mission launched by the Italian Space Agency (ASI) in April 2019. The mission is designed to collect data at global scale for a variety of applications, including those related to the cryosphere. This study presents an evaluation of PRISMA Level 1 (L1) and Level 2 (L2D) products for different snow conditions. To the aim, PRISMA data were collected at three sites: two in the Western European Alps (Torgnon and Plateau Rosa) and one in East Antarctica (Nansen Ice Shelf). PRISMA data were acquired contemporary to both field measurements and Sentinel-2 data. Simulated Top of the Atmosphere (TOA) radiance data were then compared to L1 PRISMA and Sentinel-2 TOA radiance. Bottom Of Atmosphere (BOA) reflectance from PRISMA L2D and Sentinel-2 L2A data were then evaluated by direct comparison with field data.Both TOA radiance and BOA reflectance PRISMA products were generally in good agreement with field data, showing a Mean Absolute Difference (MAD) lower than 5%. L1 PRISMA TOA radiance products resulted in higher MAD for the site of Torgnon, which features the highest topographic complexity within the investigated areas. In Plateau Rosa we obtained the best comparison between PRISMA L2D reflectance data and in situ measurements, with MAD values lower than 5 % for the 400-900nm range. The Nansen Ice Shelf instead resulted in MAD values <10% between PRISMA L2D and field data, while Sentinel-2 BOA reflectance showed higher values than other data sources.